Merge pull request #392 from sunya-ch/server-api-rebase-patch-4

feat: get type in /best-models API and pass source and type from estimator

src/kepler_model/estimate/model_server_connector.py

@@ -68,11 +68,17 @@ def make_request(power_request):
     return unpack(power_request.energy_source, output_type, response)
 
 
-def list_all_models():
+def list_all_models(energy_source=None, node_type=None):
     if not is_model_server_enabled():
         return dict()
     try:
-        response = requests.get(get_model_server_list_endpoint())
+        endpoint = get_model_server_list_endpoint()
+        params= {}
+        if energy_source:
+            params["source"] = energy_source
+        if node_type:
+           params["type"] = node_type
+        response = requests.get(endpoint, params=params)
     except Exception as err:
         print(f"cannot list model: {err}")
         return dict()

src/kepler_model/server/model_server.py

@@ -189,6 +189,7 @@ def get_available_models():
     fg = request.args.get("fg")
     ot = request.args.get("ot")
     energy_source = request.args.get("source")
+    node_type = request.args.get("type")
     filter = request.args.get("filter")
 
     try:
@@ -205,18 +206,25 @@ def get_available_models():
         if energy_source is None or "rapl" in energy_source:
             energy_source = "rapl-sysfs"
 
+        if node_type is None:
+            node_type = any_node_type
+        else:
+            node_type = int(node_type)
+
         if filter is None:
             filters = dict()
         else:
             filters = parse_filters(filter)
 
         model_names = dict()
         for output_type in output_types:
+            logger.debug(f"Searching output type {output_type}")
             model_names[output_type.name] = dict()
             for fg in valid_fgs:
+                logger.debug(f"Searching feature group {fg}")
                 valid_groupath = get_model_group_path(model_toppath, output_type, fg, energy_source, pipeline_name=pipelineName[energy_source])
                 if os.path.exists(valid_groupath):
-                    best_candidate, _ = select_best_model(None, valid_groupath, filters, energy_source)
+                    best_candidate, _ = select_best_model(None, valid_groupath, filters, energy_source, node_type=node_type)
                     if best_candidate is None:
                         continue
                     model_names[output_type.name][fg.name] = best_candidate["model_name"]

tests/estimator_model_request_test.py

@@ -40,7 +40,7 @@ def test_model_request():
     energy_source = test_energy_source
     # test getting model from server
     os.environ["MODEL_SERVER_ENABLE"] = "true"
-    available_models = list_all_models()
+    available_models = list_all_models(energy_source=energy_source)
     assert len(available_models) > 0, "must have more than one available models"
     print("Available Models:", available_models)
     for output_type_name, valid_fgs in available_models.items():

tests/weight_model_request_test.py

@@ -29,11 +29,11 @@
     os.environ["MODEL_SERVER_ENABLE"] = "true"
     energy_source = test_energy_source
 
-    available_models = list_all_models()
+    available_models = list_all_models(energy_source=energy_source)
     while len(available_models) == 0:
         time.sleep(1)
         print("wait for kepler model server response")
-        available_models = list_all_models()
+        available_models = list_all_models(energy_source=energy_source)
 
     for output_type_name, valid_fgs in available_models.items():
         output_type = ModelOutputType[output_type_name]